Hidden costs of stock fund commissions/spreads?

[ESRBob]

Did you cover this on the board already? I've been puzzling over how to find/factor in the costs my VG or DFA funds incur by trading -- direct commissions and soft dollar costs as well as bid/ask spread costs. My understanding is these costs lower overall returns of the fund below what the long-term data series show that asset class should produce. As a result, I am thinking, just like normal fund management fees, we need to subtract this off our yield assumptions (when using FireCalc, for instance).

Problem is, this data appears devilishly hard to track down -- brokerage commissions for example are hidden in obscure parts of the fund's annual income statement in the prospectus. Has anyone on this board addressed this? Do we know that for most Vanguard index funds or Tax-managed funds, that it amounts to just 0.1 or 0.2% per year (for instance) or some other reasonable rule of thumb?. A recent WSJ study showed that for a hot-money stock fund in small caps and with high turnover, it could amount to as much as 3% per year 'hidden' cost to the fund, in addition to regularly documented fees. If we owned a fund investing in a certain asset class (Smallcap International, for instance) and expected it to return us the long-term historical rates due to that asset class, we'd be toast.

This may be an old thread you've beat to death already, and if so, pls send me to links as I have not found them on my own.

thanks.

[Nords]

Speaking as a micromanaging nuclear engineer, this seems a little too far down in the weeds.

Fund managers don't want to disclose the data for their own very good (perhaps unethical if not illegal) reasons. So if we could somehow legislate disclosure, I'd still be suspicious of the data's quality.

It's difficult to tease out the individual costs of fund management, but the aggregate cost is reflected in the NAV change between 1 Jan-31 Dec vs the index's results. Breaking out the trading costs from the fund manager's "brilliance" still can't separate it from "luck". If an actively-managed fund can overcome the drag of its self-imposed trading costs, it still takes a couple decades to distinguish skill from luck. (Even Bill Miller, after 13 years of market-beating returns, is still within the limits of coin-flipping competition.) I'm not trying to reignite the active-passive controversy, but active management tends to have higher trading costs without producing a discernable yield premium. Then there's fund bloat, holding cash for redemptions, and other drags. Bill Bernstein's "Four Pillars" book has an entertaining description of Peter Lynch trapped on a hamster wheel after a few years of success at Magellan.

So for an index fund, I'd just subtract its ER (I think FIRECalc has a block for this already). A pessimistic estimate could raise that cost by 40 or 50 basis points. Conventional wisdom claims that it's not unreasonable to add another 100 basis points for every year of 100% turnover. It's not precise but it's certainly accurate enough, and it's less significant than fund loads, advisor fees, and high expense ratios.

The returns of the small-cap international index are probably easier obtained in a passive fund or an ETF that replicates the index. That has a few more complications like currency hedging, but passive indexing would lower the trading costs & spread issues for any asset class. IMO the biggest impacts will come from reducing investing expenses by avoiding them, without having to quantify them.

[unclemick]

One of my head scratchers - in lieu of data - is turnover in asset classes in the sense when you index a small class the buy/sale of stocks to 'stay in class'. I have a tiny amount of Vanguard Small Cap Value Index where the turnover runs over 20% - ? okay so what does that mean? - is there a friction cost to slice and dicing asset classes too thin? I haven't a clue - but I do think about it - sometimes.

[ats5g]

Check out this study by the Zero Alpha Group. The extra costs for Vanguard's large cap index funds appear to be very small, and even those for Windsor and Windsor II appear to be small.

- Alec

[unclemick]

2% turnover - Vanguard 500. 109% - Vanguard Small Cap Index. Now sm cap may be changing tracking index - But Bogle harps on turnover for cost as well as other reasons. Even dropping back to the 'old' 20-30% range there's a 'cost' to slicing an index class too thin. Whether it matters in a portfolio of asset classes is another matter. My INTJ mind says it can't be free.

[cute fuzzy bunny]

Hmm, I wonder if that has an effect on our running considerations of vanguard vs dfa funds. DFA clearly slices their data thinner than vanguards. So I wonder if the true returns from a 10 part DFA sliced portfolio exceed a 5-6 part vanguard portfolio...especially since the full bannana DFA port encourages foreign fixed income and microcap investments...which must have some serious trading costs and turnover.

[ESRBob]

What I read about DFA is that they try to use an "enhanced index" approach which either says "we are smart enough to not buy every single thing in the index slavishly every day to keep rebalancing, and thus we can save you money on trading costs by letting ourselves hold out a little longer before we trade. " or it says "we take a regular index and then we value-tilt it to make it pay out better". I can't quite determine which they mean, but they figure this enhances yields by a few percent in many cases. It also should make them more tax-advantaged.

http://library.dfaus.com/articles/index_enhanced_funds/

I generally don't believe in smart people and magic, but what they are doing just might work. Especially with small and international, I suspect there are shortcuts to take which will reduce turnover and enhance performance. Drift is a problem, but when we are talking about a slice of a slice index, I don't care too much -- I'd rather have the performance than perfect match to an small-cap value index (which we've just learned from Vanguard has more then one flavor one could track anyway!)